Go Back
Success Story


AI-predicted customer list generated order values 12x greater than that of rule-based segmented list

“In Senheng, we are always dedicated to deliver the best shopping experience to our customers. With this goal in mind, we’ve worked with Appier closely to help the business adapt to customer data platform and AI technology so that we can understand each customer’s behavior and preferences, whether they shop online or offline.”

Vice President

June Tai Tze Ten



Lack of transaction data for promoting new products

When it came to promoting new products, the client had no previous transactions to use as reference for which customers to target. In the particular case of a new smartphone series, the client had a difficult time gauging which type(s) of customers the product would sell to. It needed help narrowing down potentially purchasing customers so it was not blindly marketing to everyone.

Limited ability to generate list of purchasing customers to target

Like most retailers, the client relied on rule-based segmentation to figure out whom to target with what product. This traditional approach generated limited results and was in need of innovation.


An AI model with AutoML to predict target customers

Appier’s AIXON solution unified volumes of online and offline data to create an AI model with auto-machine learning that could predict target customers without transaction data. Using a product similarity funnel, Appier leveraged data from similar products to the new smartphone, such as brand, price and category, to generate a list of customers for the client. This AI-generated list generated a conversion rate that was 2.8x that of the manually-selected list.

An AI approach to segmentation

Appier’s AIXON solution took into account various parameters to predict segments of customers that would buy different products. These parameters included purchased categories, purchased brands, recency, membership, warranty and revenue. In the case of generating a user list for three distinct products–TV, fridge, and vacuum–the AI-predicted list resulted in a 30% increase in the open rate and CTR of EDMs, a 50% increase in CTR of SMS, and an overall increase of 12x in order values. These increases are all compared to the results of a rule-based segmented list.

Download case study report now